Publication: Dempster-shafer Algoritmasının Kullanımı İle Sınıflandırma Algoritmalarının Birleştirilmesi
Loading...
Files
Date
Authors
Advisor
Journal Title
Journal ISSN
Volume Title
Publisher
Fen Bilimleri Enstitüsü
Institute of Science and Technology
Institute of Science and Technology
Type
Abstract
Bu çalışmada farklı sınıflandırma algoritmalarından elde edilen sonuçların, Dempster’in Birleştirme Algoritması ile birleştirilmesinde, sınıflandırıcıların doğruluğunu da göz önünde bulundurulan bir yöntem önerilmektedir. Farklı veri takımlarıyla yapılan deneyler sonucunda, önerilen yöntemle yapılan birleştirme işleminin, birleşimde kullanılan her bir sınıflandırma algoritmasından daha doğru ve daha başarılı sonuçlar gerçeklediği gösterilmiştir.
In this study, we propose a method for improving classification results. The method consists of combining the classification results using Dempster’s Rule of Combination, considering the classifier outputs as beliefs, with the employment of degree of confidence of classifiers. In the combination we utilize some of the existing classification algorithms. We do experiments with different data sets to evaluate our proposed method and we arrive at the conclusion that our proposed method gives better classification results than each of the classification algorithms. The experiments performed on several data sets show that the employment of degree of confidence during combination results in more precise classification results.
In this study, we propose a method for improving classification results. The method consists of combining the classification results using Dempster’s Rule of Combination, considering the classifier outputs as beliefs, with the employment of degree of confidence of classifiers. In the combination we utilize some of the existing classification algorithms. We do experiments with different data sets to evaluate our proposed method and we arrive at the conclusion that our proposed method gives better classification results than each of the classification algorithms. The experiments performed on several data sets show that the employment of degree of confidence during combination results in more precise classification results.
Description
Tez (Doktora) -- İstanbul Teknik Üniversitesi, Fen Bilimleri Enstitüsü, 2005
Thesis (PhD) -- İstanbul Technical University, Institute of Science and Technology, 2005
Thesis (PhD) -- İstanbul Technical University, Institute of Science and Technology, 2005
Subject
Sınıflandırma, Dempster’in Birleştirme Algoritması, Veri Madenciliği, Veri Füzyonu, Güven Derecesi, Classification, Dempster’s Rule of Combination, Data Mining, Data Fusion, Degree of Confidence